OP25 Issues Tuning P25 P2 CC

[TNLTRPB]

I've been trying to fine tune my Nooelec NESDR Smart XTR to track the local Harris P25 P2 trunking system using boatbod's op25 fork and have just about exhausted everything I know to try (having followed Boatbod's guides and suggestions in other threads). I'm having issues getting the control channel tuned and decoding while also being able to decode the voice channels.

The best -d value I've found for getting the most tsbks (~15-20/sec) is -500 with -q of 0, but I don't get audio from the voice channels. If I let -X handle it, I get around 1-5 tsbks/sec, intermittent decodes of the voice channels, and lots of control channel timeouts in stderr.

If I adjust the tuning for best constellation clusters (around -800), I get no tsbks.

Here are the commands I'm currently using (the first entry in the .tsv is commented out so I don't have to wait for it to reacquire the CC when I restart the service):


Here are all the charts with -X specified:







If any of you SDR and op25 gurus have any suggestions on what I might try next, I'd very much appreciate it.

 

[boatbod]

Signal strength is good but your constellation plot looks terrible as does your datascope (eye) plot. If you look at the recovered symbols you can see they are all over the place rather than 4 nice straight lines so it's no surprise you're getting poor decode.

Try reducing the gain to see if that helps. Your noise floor is quite high (-70dB) but you can trade-off some of the excess signal strength for lower noise. A lower gain would also help if you have significant multipath interference.

One other possibility might be to set an offset in case the dongle has a dc spike. Try adding "-o 15000" to the command line.

 

[TNLTRPB]

Thanks! I'd tried everything except putting in the offset value. Between lowering the gain to 11 and adding the offset, it's performing much better. It's not perfect and the plots are still terrible, but it is working. I may seek an alternate site for the feed since I've already moved it into a room as far away from any other electronics as I can get it and put ferrite beads on the appropriate cables.

I wish I knew what was causing the noise floor to be so high. My 996P2 has trouble with that trunked system as well. It seems to vary with time of day, too. The 996 works great at night but struggles during the day.

 

[wgbecks]

The mixer plot in your original post doesn't look right. There shouldn't be a notch at center of the IF spectrum that may be due to the to the D.C. spike that exists with some SDR devices. Could you please make a new set of plot captures since making adjustments? I'd like to see two identical sets, one with the antenna disconnected and another with it connected.

Thank you,

Bill

 

[TNLTRPB]

Could you please make a new set of plot captures since making adjustments? I'd like to see two identical sets, one with the antenna disconnected and another with it connected.

Sure thing. Here are the post-adjustment plots with antenna attached:






And here are the FFT and mixer plots with the antenna disconnected (the constellation, data scope, and symbol plots look like random noise as you'd expect them to) and the offset/gain values the same as above:

 

[boatbod]

Even though there is a lot of noise the plots look better (more consistent) than they did in the beginning. What happens if you bring the gain up a little but leave everything else the same?

 

[TNLTRPB]

I set the gain to 14 and noticed a slight improvement, I think. It still fluctuates a fair amount due to the noise, but it's definitely getting closer to the bullseye.

 

[boatbod]

Definitely the best plots yet. Ideally you want the constellation plot to be in four tight corners, but as long as the dots stay in their own quadrants the decode should be solid.

 

[wgbecks]

Just to add, try bringing the gain up even more. I think the main issue was the D.C. spike that you seem to have corrected by
applying the offset suggested by @boatbod.

 

[TNLTRPB]

Nudging the gain up to 21 is giving me great constellation plots and ~30 tsbks/s.


I really appreciate the help you guys have given me. I'm pretty radio and tech/linux savy but this SDR stuff is an entirely new beast. Special thanks to @boatbod for the guide he wrote up in this thread.

One more question: what's the best way to handle rotating control channels? Do I go with the first approach I tried of listing them all out in the .tsv file?

 

[wgbecks]

Looking great! Just populate your trunk.tsv control channel section with a comma separated list and you'll be golden.

 

[boatbod]

Looking good now!

 

[maus92]

Definite improvement. How is your sdr attached to your computer? You *could* remote the sdr by using an USB extension cable, and move the antenna away somewhat farther from the computer to see if that is the source of your noise.

 

[TNLTRPB]

Definite improvement. How is your sdr attached to your computer? You *could* remote the sdr by using an USB extension cable, and move the antenna away somewhat farther from the computer to see if that is the source of your noise.

Right now, the SDR is plugged directly into the pi with the telescopic antenna set to 1/4λ at 855MHz ~6' away from the pi and using an old steel baking pan as a ground plane. I do want to move both the SDR and antenna away from the pi a bit but that will have to wait until I'm back from vacation.